[sdiy] really inaccurate zeners

Oren Leavitt obl64 at ix.netcom.com
Wed Jul 1 01:06:25 CEST 2020

I'll fifth this!

Reverse biased transistor BE junctions work great when you need a zener 
with a sharp knee at low (< 1mA) currents. The "zener" voltage will be 
about 5+ volts or so. Power zeners are very "mushy" at low currents.

- Oren

On 6/30/20 5:04 PM, Chris McDowell wrote:
> I'll just second / third / fourth this. We run into this at work often, and are now in the habit of always checking the stated reverse leakage current of zeners (or just look at the graph of Iz vs Vz  if they have one). We are usually pretty stingy with wasted mW, so have to track down very low leakage zeners if we ever need them between 3V and 5V. Maybe it's easier to fabricate tighter tolerance and low leakage zeners at higher voltages, someone with more experience there could confirm or deny this, but the low voltage zeners tend to have way worse leakage.
> Chris
>> On Jun 30, 2020, at 4:23 PM, Neil Johnson <neil.johnson71 at gmail.com> wrote:
>> Hi David,
>>> The amplitude control is achieved in each core with a single zener diode inside a "zener bridge" consisting of four 1N4148s.  Since the 4148s drop approximately 0.5V, a 3.9V zener (1N4370) should give a total drop of roughly 5V, which should guarantee waveforms which are 10Vpp.
>> Umm, I thought the 1N4370 was a 2.4V device?  All the datasheets I can
>> find confirm this.
>>> One was about right, but the other was way low.  I replaced the zener of the low one, and it got worse.  So then I decided to actually measure some zeners.  After much grumbling, the bottom line is that I had to delve into my bag of 1N4372s (4.7V) before I found two zeners that read almost exactly 3.9V.  They were all very low relative to their nominal values.  (I tested them by putting a 3.3k resistor between +15V and the cathode, and connecting the anode to ground.  I use 3.3k resistors on the comparator in my VCO core.)
>> Umm, and 1N4372 are supposed to be 3.0V.
>> You may want to recheck the part numbers?
>>> So now I have decided that I have to measure each and every zener diode, and separate them into plastic bags based on their actual voltages rather than their denominations.  In this way, after a certain amount of tedium, I will have a collection of precision zener diodes (within +/- 0.05V).  However, I shouldn't have to do this.  I would expect zeners to be off by 0.2 (or, at most, 0.3V) from their nominal values, but to cross over one or two (or even three) denominations?  That is unacceptable.  What are these zener diode manufacturers doing wrong?
>> Well, zeners - like all electronic components - have tolerances.  In
>> your case, assuming you have 1N4372 diodes, without any suffix they
>> have a 10% tolerance.  You can get tighter tolerance parts but you pay
>> more for them (they sort them on the production line).
>> Also, as Tom pointed out, the advertised voltage is at a stated
>> current (Izt - Zener test current). For the 1N437x series this is
>> 20mA.  At lower test currents you will measure a lower reverse
>> voltage.  For example, if you dig out the Motorola datasheet on these
>> parts:
>> https://datasheetspdf.com/pdf-file/1388706/Motorola/1N4370/1
>> on page 6 you can see how much the zener voltage varies with current
>> (answer: quite a lot!)
>> Neil
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at synth-diy.org
>> http://synth-diy.org/mailman/listinfo/synth-diy
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at synth-diy.org
> http://synth-diy.org/mailman/listinfo/synth-diy

More information about the Synth-diy mailing list